1. Field of the Invention
The present invention relates to a reproducing apparatus for reproducing image signals, and more particularly, to an image reproducing apparatus for reproducing image signals by decoding high-efficiency coded image signals and by concealing the data which cannot be decoded.
2. Description of the Related Art
An example of a conventional image reproducing apparatus of the above-described is a digital VTR (video tape recorder) which digitizes an image signal and records or reproduces the digital image signal. The digital VTR employs the band compression coding technique to compress and record or reproduce image data.
Orthogonal transform coding is known as a technique for high-efficiency coding of image signals. In a method employing orthogonal transform coding, after an image signal is divided into blocks each consisting of a plurality of pixels, orthogonal transform coding, such as discrete cosine transform (hereinafter referred to as DCT), is performed on each of the blocks. The obtained coefficients of transform are quantized or entropy coded.
Since coding is performed for each block in the above-described coding method, when errors occur in the coded data during the reproduction of an image signal, coding errors are generated that affect entire blocks. Further, since variable-length codes are frequently used, a decoding error spreads over a plurality of blocks, thus enlarging a deterioration in the image quality. In the case of such a decoding error, if there are data that cannot be corrected by the error correction process, these data are interpolated, as non-correctable error data (hereinafter referred to as error data), using the data of a block located close thereto in terms of time and space.
To interpolate the image data of a block (hereinafter referred to as an error block) containing such error data, conventionally it has often been practiced to selectively employ inter-frame interpolation and intra-field interpolation. That is, whether there is a moving image in an error block is determined by obtaining an inter-frame correlation (which is the result of the intra-block addition of differential absolute values between the pixels of the present frame and the pixels of the preceding frame) of the image data of blocks above and below a drop-out (i.e., error) block. If there is a moving image, intra-field is selected. If there is no moving image, interframe interpolation is selected.
However, in the above-described type of interpolation, if there is an error in determining as to whether there is a moving image in the error block, i.e., if a moving image is erroneously not detected, interpolation is performed using pixel data which is less correlated to the pixel block to be interpolated, thus deteriorating the quality of a reproduced image.